This proposal builds upon the on-going multi-center family study entitled International Study of Familial Glioma (Gliogene - R01CA119215), and the funded Brain SPORE at University of California, San Francisco (UCSF) to explore genetic predictors of neurocognitive (NC) and clinical outcome. We will leverage these funded resources by using the existing infrastructures for recruiting glioma patients at The University of Texas . D. Anderson Cancer Center (UTMDACC) and UCSF. We will recruit from both sites a total of 800 patients between the ages of 18 and 70 years with newly diagnosed high-grade glioma (GBM). NC outcomes are shown to vary across patients and such variation may be partly due to genetic differences among patients that could potentially modulate treatment response, disease progression, and NC sequelae. We therefore propose a multidisciplinary integrative approach using a panel of biomarkers in patients receiving standardized treatment protocols to determine factors related to NC function, symptom burden and clinical outcomes (such as treatment toxicity and survival) all of which affect overall QOL. Hypothesis: Genetic polymorphisms in NC, metabolizing genes, DMA repair pathway, and inflammation genes alter the NC and clinical outcome of brain tumor patients and, in turn, impact their QOL. Specific Aim 1: To recruit approximately 100 newly diagnosed GBM patients per year from the neurooncology clinics at UTMDACC and UCSF. From these patients' DNA, we will genotype a panel of NC, metabolizing, DNA repair and inflammation genes. Specific Aim 2: To conduct NC evaluations on all patients treated and followed at each institution (200/year N=800 over 4 years) following the standardized testing established for a number of collaborative group and pharmaceutical company sponsored studies. Specific Aim 3: To integrate data derived from Aims 1 and 2 to determine if variants in the NC, metabolizing, DNA repair, and inflammation pathway genes alter overall QOL. Using the wealth of clinical information abstracted from medical records and data generated from this proposal, we will develop a Risk Prediction Model to determine which patients are more likely to develop NC dysfunction during their disease course and could therefore be targeted for personalized treatments that would potentially impact their QOL. Brain tumors cause progressive NC deficits similar to other neurodegenerative diseases, and these deficits frequently worsen because of treatment effects. Some patients even develop treatment related dementia that leads to death without evidence of tumor progression. Thus, the ability to determine markers that can identify those vulnerable patients would be of tremendous benefit and allow for alterations of primary treatment or neuroprotective therapies to reduce the risk of brain tumor therapy.